Understanding Electrochemically Induced Olefin Complexation: Towards Electrochemical Olefin-Paraffin Separations

Abstract

Olefin-paraffin separation is a critical yet energy-intensive process in the chemical industry, accounting for over 250 trillion BTU/year of global energy consumption. This work explores the use of redox-active nickel maleonitriledithiolate complex for olefin-paraffin separations. Key performance factors, namely the electrochemical oxidation of the complex and the olefin capture utilization fraction, were systematically quantified. Electrochemical studies revealed near-complete oxidation of Ni(II) to Ni(IV) species, suggesting that the electrochemical oxidation step is not a limiting factor in olefin capture. The utilization fraction was found to be strongly dependent on the complexation equilibrium behavior between olefin-bound and unbound state of the complex. Time-resolved kinetic measurements unveiled a sluggish complexation rate, requiring over 36 hours to approach equilibrium. These insights highlight the importance of driving the complexation equilibrium and improving the kinetics to enhance the performance of Ni-based electrochemical swing absorbers for energy-efficient olefin-paraffin separations. The findings lay the groundwork for future optimization strategies and industrial implementation of this sustainable separation technology.

1. Introduction

Light olefins are key building blocks for chemical manufacturing and are used to produce many chemical commodities, including polymers and their precursors. Global production of ethylene and propylene was more than 220 and 150 million metric tons per year, respectively, in 2022.^{1, 2} Current ethylene and propylene production involves separating the olefin and paraffin gas mixtures emerging from thermal cracking reactors. Purifying ethylene and propylene is technologically challenging due to their similar molecular structures and physicochemical properties to that of their paraffin counterparts, and involves energy intensive cryogenic distillations, which require at least 3.6 MJ/kg olefin separated.^{3, 4} Enabling alternative, electrified, non-thermal olefin-paraffin separations has the potential to improve energy efficiency and to integrate renewable electricity sources into the separation process.

Various non-thermal separation methods have been explored for purifying olefins, but challenges persist in scaling up and deploying these techniques. Membrane-based separations show promise, potentially reducing energy requirements by up to 85% compared to cryogenic distillation.^{5, 6} Membranes that rely on a solution-diffusion mechanism face limitations due to the similarity in size and dipole moments of light olefins and paraffins.^7–12 Molecular-sieving methods that use carefully engineered metal-organic frameworks (MOF) or zeolites with angstrom precision have been shown to improve selectivity.^{7, 8, 13–17} Alternatively, metallic species like silver(I) ions, known for forming weak pi-bond complexes with alkenes, can be embedded in membranes and enhance separation via a facilitated-transport mechanism.^{8, 13, 18–24} However, these methods require pre-treatment of the olefin stream to avoid reactions between silver ions and impurities such as hydrogen sulfide or acetylene.^25–27

Electrochemically modulated separation systems are another promising alternative to traditional separation processes. For example, Cu(I) ions can selectively capture olefins through pi-bond complexation and release the captured olefin when electrochemically oxidized to Cu(II). The ability to reversibly modulate the oxidation state of the Cu ions by controlling the electrode potential allows olefin separations to occur via this capture-and-release mechanism.^{28, 29} Wang and Stiefel demonstrated that nickel dithiolene complexes could be used in electrochemically modulated olefin/paraffin separations by binding olefins through their ligands rather than the metal center, and in this way avoid poisoning with impurities in the gas stream.³⁰ Similar concepts could be deployed in a solid-state porous coordination polymers (PCP) with integrated metal bis(dithiolene) or olefin-attracting MOFs that act as the active redox site.^31–35 More recently, our group demonstrated an electrochemical swing absorber where redox-active nickel dithiolene complexes were integrated into a solid-state membrane-electrode assembly (MEA) device, albeit with low separation performance [i.e., low apparent Faradaic efficiency (FE), < 0.4%, and utilization fraction (UF), < 1.5%].³⁶ This demonstration used [Ni(mnt)₂]ⁿ complexes (mnt = [S₂C₂(CN)₂]^2-, maleonitriledithiolate; n = 0, 1-, 2-), where the nickel’s oxidation state (OS) (IV, III, or II) was electrochemically modulated. Figure 1 depicts the reversible redox cycle where different oxidation states of [Ni(mnt)₂]ⁿ species can be electrochemically accessed. One state has a high affinity to complex olefins (n = 0, OS = IV), while the other state has low affinity towards them (n= 2-, OS = II), enabling electrochemically modulated separation. To improve the performance of electrochemical swing absorbers in terms of energy sustainability it is necessary to fundamentally understand the key factors that impact FE and UF, even though the MEA device demonstrated near-100% chemisorption selectivity in olefin capture-and-release.³⁶ Improvement of FE and UF can potentially make this separation approach less energy intensive than traditional distillation methods, paving a way towards sustainable separation process for olefin purifications.

Figure 1.

Electrochemical H-Cell Reactor is depicted on the left. Electrochemical Redox Cycle utilizing [Ni(mnt)₂]ⁿ in a capture-and-release of 1-pentene olefin is shown on the right. The efficiency of the steps shown in the redox cycle can grant a more thorough understanding of the electrochemical effect on olefin separation.

In this work, we systematically explore the impact of electrochemical conditions and the olefin capture environment on separation performance in solutions containing [Ni(OS)(mnt)₂]ⁿ complexes and 1-pentene (C₅H₁₀) as a model olefin to understand the origin of key performance limitations. 1-pentene was used as a model liquid olefin to allow quantitative control of olefin concentration in the reaction solution environment. Since the MEA experiment from the past study thoroughly demonstrated the ability of [Ni(OS)(mnt)₂]ⁿ to facilitate propylene-propane separation³⁶ and because paraffin only plays a role as a spectator, it is assumed that the presence of paraffin counterpart is not necessary to help observe what limiting phenomena affect the efficiency of the olefin capture-and-release mechanism. In our study, we implemented electrochemical methods together with UV-Vis and Nuclear Magnetic Resonance (NMR) spectroscopic techniques to quantify the oxidative conversion fraction of [Ni(II)(mnt)₂]²⁻ and the UF of [Ni(IV)(mnt)₂]⁰ species. Coupling an understanding of the electrochemical conversion of [Ni(II)(mnt)₂]²⁻ and its UF through simultaneous UV-Vis and H-NMR analysis can enhance our understanding of factors that drive olefin complexation efficiency under varying electrochemical conditions.

2. Experimental Methodology

To uncover the relationship between the electrochemical conditions and the efficiency of oxidation-induced capture of olefins, 3 key experimental steps, shown in Figure 2, were used to quantify the electrochemical conversion of [Ni(II)(mnt)₂]²⁻ and its UF under varying reactant concentrations and voltametric parameters. In Figure 2A, constant equivalent charge, depending on the initial concentration of [Ni(II)(mnt)₂]²⁻, is applied in each experiment to allow consistent comparison of its conversion. The equivalent charge, Q_equivalent, is calculated as the theoretical minimum total charge needed to oxidize all of the [Ni(II)(mnt)₂]²⁻ species to [Ni(IV)(mnt)₂]⁰ present in the reaction solution (i.e., Q_equivalent = n_oxidation ∙ F ∙ moles of [Ni(II) (mnt)₂]²⁻, where n_oxidation = 2 electrons and F = Faraday’s Constant). By controlling the reaction time, t_rxn, and maintaining constant current, I_constant, through chronopotentiometry, the total applied charge can be adjusted to the target equivalent charge in each experiment based on the specified starting reactant concentration and current density.

Figure 2.

(A) Applied Equivalent Charge, Q_equivalent, is calculated before each experimental trial to keep consistent charge ratio vs. the initial concentration of [Ni(II)(mnt)₂]²⁻. Charge applied to the electrochemical cell is the theoretical minimum required to oxidize all [Ni(II)(mnt)₂]²⁻ into [Ni(IV)(mnt)₂]⁰. The applied charge is controlled by the reaction time, t_rxn, under constant current, I_constant. (B) UV-Vis Quantification of Conversion Rate of [Ni(II)(mnt)₂]²⁻ is made possible because each oxidation state of [Ni(mnt)₂]ⁿ has its own distinct set of band energy wavelength. The peaks, highlighted at 4 wavelengths shown above, decrease in height magnitude as more [Ni(II)(mnt)₂]²⁻ is oxidized. By tracking the concentration of [Ni(II)(mnt)₂]²⁻ from peak height in UV-Vis spectroscopy, the conversion rate can be determined. The peak at 310 nm was deliberately not used for tracking the [Ni(II)(mnt)₂]²⁻ concentration because not only does it overlap with the spectra feature of [Ni(III)(mnt)₂]⁻, but it also overlaps with that of the olefin complex. (C) H-NMR Quantification of [Ni(IV)(mnt)₂]⁰∙C₅H₁₀ leverages the unique multiplicity peak (i.e., quartet of doublet) of the hydrogens attached to the first carbon of 1-pentene bonded to the sulfur atoms of the [Ni(IV)(mnt)₂]⁰∙C₅H₁₀. UF can then be calculated using the equation shown above.

To quantify the conversion, the concentration of the oxidized species, [Ni(II)(mnt)₂]²⁻, is determined via UV-Vis spectroscopy by tracking changes in the absorption spectra, as shown in Figure 2B. All UV-Vis measurements were performed immediately after application of an oxidative current to the reaction solution. In the process of electrochemical oxidation [Ni(III)(mnt)₂]^1- and [Ni(IV)(mnt)₂]⁰ are converted from [Ni(II)(mnt)₂]²⁻ and, as oxidation progresses, it is assumed that [Ni(III)(mnt)₂]^1- is also converted into [Ni(IV)(mnt)₂]⁰. Since [Ni(II)(mnt)₂]²⁻, [Ni(III)(mnt)₂]^1-, and [Ni(IV)(mnt)₂]⁰ each have characteristic absorption features, the individual concentration of each species can be quantified. [Ni(II)(mnt)₂]²⁻ and [Ni(III)(mnt)₂]^1- display prominent energy bands at 270, 310, 385 and 470 nm, and at 310 and 860 nm wavelength, respectively. The lower absorption peaks at 385 and 470 nm wavelength for [Ni(II)(mnt)₂]²⁻ and at 860 nm wavelength for [Ni(II)(mnt)₂]^1- are consistent with known UV-Vis characteristics for this complex.^{37, 38} [Ni(IV)(mnt)₂]⁰ did not exhibit any prominent absorption feature in the UV-Vis spectra. Wang et al. showed that an absorbance peak grows in magnitude at the energy band wavelength towards the ultraviolet side of the spectrum as [Ni(IV)(mnt)₂]⁰ species couple with olefins.³⁰ This absorption feature appearance which correlates with the formation of the 1-pentene adduct was experimentally observed at 310 nm. In the UV-Vis spectra shown in Figure 2B the peaks for Ni(II) and Ni(III) have diminished, demonstrating complete oxidation of all [Ni(II)(mnt)₂]²⁻.

Quantifying the complexation of olefins with [Ni(IV)(mnt)₂]⁰ is challenging due to overlapping UV-Vis absorption features from multiple chemical species. To overcome this challenge and to provide insights into limiting factors in the electrochemical separation process, we performed H-NMR analysis to quantitatively study the complexation of [Ni(IV)(mnt)₂]⁰ with 1-pentene, forming [Ni(IV)(mnt)₂]⁰·C₅H₁₀. High-frequency NMR allowed us to observe detailed molecular features from the complexation process even for small concentrations of [Ni(IV)(mnt)₂]⁰·C₅H₁₀. It should be noted that the [Ni(IV)(mnt)₂]⁰ species do not contain any hydrogen atoms and are undetectable on H-NMR spectra. The complex, however, can be quantified because the coupled 1-pentene olefin molecule contains characteristic hydrogen atoms at distinct chemical shifts that are different than those from the free olefin. Unique chemical shifts were located at 4.75 and 3.68 ppm, belonging to the hydrogen atoms bonded to the first and second carbon of the complexed 1-pentene, respectively. Peak multiplicities for these 2 groups of hydrogen atoms closest to the sulfur atoms on [Ni(IV)(mnt)₂]⁰ were verified on experimental NMR spectra, details of which can be found in Section S1 in the supporting information. The ratio of integrated peak areas between these two sets of multiplet peaks was approximately 2, which is in agreement with the theoretical ratio because there are 2 hydrogen atoms on the first carbon atom for every 1 hydrogen atom on the second carbon atom. These correlations further reinforced the observation that at the specified chemical shifts, these peaks seen in the NMR spectra are unique to [Ni(IV)(mnt)₂]⁰·C₅H₁₀. For concentration quantification, peak area integration of the quartet of doublet peaks at 4.74 ppm was used throughout this study, as shown in Figure 2C, which provided the most accurate quantification. The H-NMR measurement was typically taken 72 hours after the end of the oxidation reaction to ensure that the equilibrium state of the complexation is reached (see Section 3 for additional details on the calculation of equilibrium constant).

3. Results and Discussion

3.1. Understanding the electrochemical conversion of [Ni(II)(mnt)₂]²⁻

Throughout this work, the oxidation of [Ni(II)(mnt)₂]²⁻ behavior was found to be relatively consistent, even under varying electrochemical conditions. The complete disappearance of the observed absorption peaks, specifically at 270, 385, 470, and 860 nm, shown in Figure 2B was observed in all of our oxidation experiments regardless of initial [Ni(II)(mnt)₂]²⁻ or 1-pentene concentration as demonstrated in Figure 3. These results suggest that the oxidation of [Ni(II)(mnt)₂]²⁻ is highly efficient, and that the electrochemical OS manipulation of [Ni(OS)(mnt)₂]ⁿ is unlikely to be a limiting factor in the olefin-capture cycle.

Figure 3.

(A) UV-Vis Absorbance over Varying [C₅H₁₀] shows the absorbance effect from 1-pentene concentration at constant 5 mM [Ni(mnt)₂]ⁿ concentration and 0.57 mA/cm² oxidative current density. Major peaks belonging exclusively to Ni(IV) and Ni(III) have diminished during electrochemical oxidation. At band wavelength of 310 nm, the absorbance peak appears to increase with higher 1-pentene concentration environment. (B) UV-Vis Absorbance over Varying [[Ni(mnt)₂]ⁿ] shows the absorbance effect from [Ni(mnt)₂]ⁿ concentration at constant 100 mM C₅H₁₀ and 0.57 mA/cm² oxidative current density. Similarly, major peaks belonging exclusively to Ni(IV) and Ni(III) have diminished during electrochemical oxidation. However, at the wavelength of 310 nm, absorbance peak appears to decrease with higher [Ni(mnt)₂]ⁿ concentration. The absorbance peak at 310 nm wavelength is directly related to the mole fraction of [Ni(mnt)₂]ⁿ·C₅H₁₀ affected by the equilibrium between bound and unbound [Ni(mnt)₂]ⁿ species.

In Figure 3A, a positive correlation can be observed between 1-pentene concentration and the magnitude of the 310 nm absorption peak. This suggests that the 310 nm peak is not just a feature of the free [Ni(II)(mnt)₂]²⁻ and [Ni(III)(mnt)₂]⁻, but also a characteristic associated with the formation of [Ni(IV)(mnt)₂]⁰·C₅H₁₀. These UV-Vis results correlate well with H-NMR observations (discussed in the subsequent subsections) where higher concentrations of [Ni(IV)(mnt)₂]⁰·C₅H₁₀ are observed under richer initial olefin environments. This direct relationship is further supported by the negative correlation between the initial [Ni(II)(mnt)₂]²⁻ concentration and the peak magnitude at 310 nm band wavelength depicted in Figure 3B. As the mole fraction of equilibrated [Ni(IV)(mnt)₂]⁰·C₅H₁₀ complex decreases at higher [Ni(II)(mnt)₂]²⁻ concentration, the absorbance peak at 310 nm weakens in magnitude. These UV-Vis spectroscopy characteristics arising from the complex formation are consistent with those reported by Wang et al.³⁰

3.2. Understanding the utilization of [Ni(II)(mnt)₂]²⁻ in 1-pentene complexation

The concentration of 1-pentene in the vicinity of [Ni(IV)(mnt)₂]⁰ species is expected to impact the rate of complexation and thus be a determining factor in the UF of the complex. Our results, shown in Figure 4, demonstrate that the UF increases with olefin concentration and reaches a value of approximately 50% at initial concentrations of 100 mM 1-pentene and 5 mM [Ni(II)(mnt)₂]²⁻. This UF behavior is consistent with an equilibrium between the free state of the olefin and its bound state,

$${{\text{[Ni(IV)(mnt)}}_{\text{2}}\text{]}}^{\text{0}}{\text{+C}}_{\text{5}}{\text{H}}_{10}\rightleftarrows {{\text{[Ni(IV)(mnt)}}_{\text{2}}\text{]}}^{\text{0}}\cdot {\text{C}}_5{\text{H}}_{10}$$ Eq. 1

which determines the final [Ni(IV)(mnt)₂]⁰·C₅H₁₀ concentration. Using an experimentally determined complexation reaction equilibrium constant of K_eq = 10.2 M⁻¹, we demonstrate that the observed UF can be largely explained by an equilibrium behavior as shown in Figure 4B.

Figure 4.

(A) H-NMR of [Ni(mnt)₂]⁰·C₅H₁₀ over varying [C₅H₁₀] shows the H-NMR spectrum segment of the quartet of doublet peak of the resulting complex over varying 1-pentene concentration at constant 5 mM [Ni(mnt)₂]²⁻ initial concentration and 0.57 mA/cm² oxidative current density. The initial red plot represents the “control” where no oxidative current is applied. Note that the peaks increase in height and integral area as more olefin is present in the solution medium. (B) UF over varying [C₅H₁₀] shows experimental UF of the converted [Ni(II)(mnt)₂]²⁻ that coupled with 1-pentene over varying 1-pentene concentration at constant 5 mM [Ni(mnt)₂]ⁿ initial concentration and 0.57 mA/cm² oxidative current density. Error bars represent standard deviations of triplicate trials. The red dotted line represents the theoretical equilibrium UF assuming K_eq = 10.2 M⁻¹, as found in section 3.3. The amount of 1-pentene captured ranged from approximately 0 to 50 μmoles.

$$K_{eq}=\frac{\left[{\left[\text{Ni}\left(\text{IV}\right){\left(\text{mnt}\right)}_2\right]}^0\cdot {\text{C}}_5{\text{H}}_{10}\right]}{\left[{\left[\text{Ni}\left(\text{IV}\right){\left(\text{mnt}\right)}_2\right]}^0\right]\left[{\text{C}}_5{\text{H}}_{10}\right]}$$ Eq. 2

The effect of changing [Ni(II)(mnt)₂]²⁻ concentration is expected to follow a similar equilibrium behavior as the one observed for varying 1-pentene concentrations. Figure 5 shows the experimentally measured UF with respect to initial concentration of [Ni(II)(mnt)₂]²⁻ at a constant 100 mM 1-pentene concentration. As observed from our results, the concentration of the [Ni(IV)(mnt)₂]⁰·C₅H₁₀ monotonically increased as we increased the initial concentration of [Ni(II)(mnt)₂]²⁻ (Figure 5B). However, despite the increase of [Ni(IV)(mnt)₂]⁰·C₅H₁₀ complex concentration, both the experimental and theoretical UFs, as shown in Figure 5C, exhibited a gradual decline at higher initial concentration of [Ni(II)(mnt)₂]²⁻. This decrease is consistent with the reaction equilibrium described above. These results demonstrate that the UF has a weak dependence on the initial concentration of [Ni(II)(mnt)₂]²⁻, especially in the small window of accessible concentrations given its low solubility.

Figure 5.

(A) H-NMR of [Ni(mnt)₂]⁰·C₅H₁₀ over Varying [[Ni(mnt)₂]ⁿ] shows the H-NMR spectrum segment of the quartet of doublet peak of the resulting complex over varying [Ni(mnt)₂]ⁿ concentration at constant 100 mM 1-pentene initial concentration and 0.57 mA/cm² oxidative current density. The initial red plot represents the “control” where no oxidative current is applied. Note that the peaks increase in height and integral area as more olefin is present in the solution medium. (B) [[Ni(mnt)₂]⁰·C₅H₁₀] over varying [[Ni(mnt)₂]ⁿ] shows the positive correlation between the complex concentration and the initial [Ni(mnt)₂]ⁿ concentration, derived from H-NMR analysis. (C) UF over varying [[Ni(mnt)₂]ⁿ] shows experimental UF of the converted [Ni(II)(mnt)₂]²⁻ that coupled with 1-pentene over varying [Ni(mnt)₂]ⁿ concentration at constant 100 mM 1-pentene concentration and 0.57 mA/cm² oxidative current density. The red dotted line represents the theoretical equilibrium UF assuming K_eq = 10.2 M⁻¹, as found in section 3.3. Error bars represent standard deviations of triplicate trials. The amount of 1-pentene captured ranged from approximately 0 to 140 μmoles.

In order to understand the impact of electrochemical conditions on the complexation behavior, we performed experiments where the current density was varied between 0.11 to 0.80 mA/cm² while maintaining a constant initial concentration of 5 mM and 100 mM for [Ni(II)(mnt)₂]²⁻ and 1-pentene, respectively. Our results demonstrated that current density has no effect on the UF (Figure 6). Current density was determined to play no significant role even at varying initial concentrations of [Ni(II)(mnt)₂]²⁻ (see Section S6 in supporting information). This observation suggests that near complete oxidation of [Ni(II)(mnt)₂]²⁻ can be achieved, regardless of current density, and that UF is primarily driven by the equilibrium reaction between the Ni complex and 1-pentene.

Figure 6.

(A) H-NMR of [Ni(mnt)₂]⁰·C₅H₁₀ over Varying Current Density shows the H-NMR spectrum segment of the quartet of doublet peak of the resulting complex over varying current density at constant initial concentrations of 5 mM [Ni(mnt)₂]²⁻ and 100 mM 1-pentene. The initial red trace represents a control experiment where no oxidative current is applied. Note that all of the blue-colored post-oxidation peaks have similar integration area because current density did not have an effect on the complexation equilibrium. (B) UF over Varying Current Density shows UF of the converted [Ni(II)(mnt)₂]²⁻ that coupled with 1-pentene over varying current density, derived from H-NMR analysis. The red dotted line is a best-fit horizontal line, signifying a zeroth order relationship between applied current density and final equilibrated concentration of [Ni(mnt)₂]⁰·C₅H₁₀. Error bars represent standard deviations of triplicate trials.

It is important to note that UF is most significantly tuned by the concentration of olefin in the reaction environment rather than by the concentration of [Ni(mnt)₂]ⁿ in all experimental trials depicted in Figures 4, 5, and 6. [Ni(mnt)₂]ⁿ salt precipitates out of the acetonitrile solution after surpassing 20 mM concentration, while 1-pentene stays highly miscible with reaction solution, indicative of the relatively limited range of workable concentrations with [Ni(mnt)₂]ⁿ. Studying the effect of varying the concentration of [Ni(mnt)₂]ⁿ, as shown in Figure 5, served mainly for the purpose of determining whether the complexation equilibrium played a significant role in affecting the final complexation rate.

3.3. Understanding the complexation reaction kinetics

The results presented in the previous subsections suggest that the formation of the [Ni(IV)(mnt)₂]⁰·C₅H₁₀ complex is driven by an equilibrium reaction between [Ni(IV)(mnt)₂]⁰ and 1-pentene, and that several hours were required to achieve this equilibrium. To better understand the reaction kinetics of the complexation behavior and provide insights for its possible implementation in olefin-paraffin separations, we performed a series of experiments where the formation of [Ni(IV)(mnt)₂]⁰·C₅H₁₀ was tracked over time via time-resolved NMR analysis (Figure 7 and Figure S3 in supporting information). Given that our UV-Vis analysis (Section 3.1) demonstrated that the process was not limited by the electrochemical oxidation of [Ni(II)(mnt)₂]²⁻ and that the rate of this step could be effectively controlled by adjusting the current density, we proceeded to study the complexation of [Ni(IV)(mnt)₂]⁰ with 1-pentene independently of the electrochemical step. To that end, we performed experiments where [Ni(II)(mnt)₂]²⁻ was oxidized in a saturated LiNO₃ solution in acetonitrile without the presence of olefins, and then subsequently we added 1-pentene and performed time-resolved H-NMR experiments over the course of the complexation. This experimental approach allowed us to control the initial reactant concentrations of the Ni complex and 1-pentene, while accurately determining the time of reaction. Our results show that the concentration of [Ni(IV)(mnt)₂]⁰·C₅H₁₀ monotonically increases over 72 hours of reaction, with the largest changes observed for the first 36 hours and then following an asymptotic trend towards its equilibrium value. Using the results from samples equilibrated for > 72 hours (see Section S4 in supporting information) and assuming that the converted [Ni(II)(mnt)₂]²⁻ species all became [Ni(IV)(mnt)₂]⁰, we determined K_eq to be approximately 10.2 M⁻¹ (with standard deviation of 2.61 M⁻¹). These results underscore the slow complexation speed which can limit the implementation of [Ni(OS)(mnt)₂]ⁿ complex for olefin separations, regardless of the fast rate for electrochemical oxidation. Allowing sufficient complexation time is a vital factor to achieve high degrees of coupling between [Ni(IV)(mnt)₂]⁰ and 1-pentene, and this slow process has likely limited the performance of previous implementations of [Ni(OS)(mnt)₂]ⁿ in electrochemical olefin/paraffin separation devices.³⁶

Figure 7.

Complex Concentration Profile shows the concentration of [Ni(IV)(mnt)₂]⁰·C₅H₁₀ complex gradually reaching an asymptotic slope representing equilibrium state of the reaction. The equilibrium constant, K_eq, can be determined for all experiments done in Section 2 because 72 hours was given for the reactant species to complex, more than enough time to reach equilibrium condition. The above concentration profile was observed from a time-dynamic experiment with initial concentrations of 4 mM [Ni(mnt)₂]ⁿ and 8 mM 1-pentene.

3.4. Decomplexation of 1-pentene from [Ni(IV)(mnt)₂]⁰·C₅H₁₀

Limited experimental trials were conducted to demonstrate decomplexation of 1-pentene from [Ni(IV)(mnt)₂]⁰·C₅H₁₀, details of which can be found in Section S5 in supporting information. The appearance and disappearance of unique quartet of doublet peak occurred in positive correlation to the electrochemical oxidation and reduction of the [Ni(OS)(mnt)₂]ⁿ species, respectively. This demonstration depicts the ability to quantitatively observe both the complexation and decomplexation of the olefin molecule in the overall capture-and-release mechanism shown in Figure 1. Furthermore, this reduction attempt revealed the technical difficulties in characterizing the life span of separation systems employing electrochemically induced capture and release of olefins because of the necessity to accommodate the long complexation equilibration time per cycle. A possible approach to making this separation process practical for widespread application would not only include optimized FE and UF, but also ways to decrease the equilibration time to increase the throughput rate of olefin capture. This technical obstacle regarding equilibration time warrants future efforts to improve the industrial feasibility of this separation process.

4. Conclusion

This study explored the impact of electrochemical and reaction conditions in the complexation of [Ni(OS)(mnt)₂]ⁿ with olefins, providing insights into the factors that drive performance for the implementation of these organometallic complexes in electrochemical olefin-paraffin separations. Our study showed that electrochemical oxidation of [Ni(II)(mnt)₂]²⁻ proceeded to completion with high efficiency and independently of applied current density. These results suggest that the electrochemical oxidation step does not play a limiting role in the olefin-capture mechanism. While exploring the impact of varying concentrations of 1-pentene and [Ni(OS)(mnt)₂]ⁿ on the reaction, we found that the UF is primarily driven by the equilibrium between the bound and unbound state of the olefin with the Ni complex. Furthermore, time-resolved kinetic studies confirmed that the coupling rate between [Ni(IV)(mnt)₂]⁰ and 1-pentene is slow, and that achieving near-equilibrium concentrations requires >36 hours of reaction. These findings demonstrate the need to provide sufficient complexation time to increase the UF of [Ni(OS)(mnt)₂]ⁿ in electrochemical olefin separation devices. The insights provided by this study suggest that molecular design, reaction engineering, and reactor design strategies aimed at improving the olefin complexation kinetics or at driving the equilibrium towards the formation of [Ni(IV)(mnt)₂]⁰·C₅H₁₀ could improve the potential of Ni complexes for industrial olefin-paraffin separations.

6. Detailed Experimental Methods

Electrochemical oxidation in this study was facilitated by a custom electrochemical H-cell reactor (a detailed figure can be found in Section S2 in supporting information). The reactor was built by machining Teflon blocks purchased from McMaster-Carr into specific reactor module shapes designed in Autodesk Inventor. Tetrabutylammonium Nickel (II) Maleonitriledithiolate was acquired from Tokyo Chemical Industries at a purity of approximately 97%. HPLC-grade acetonitrile solvent, 1-pentene 98%, and 99.9% lithium nitrate powder used to saturate the acetonitrile solvent were acquired from Sigma-Aldrich. Platinum mesh and wire used on the counter-side were also acquired from Sigma-Aldrich. The apparent area of the mesh used as a counter electrode was approximately 1 cm². Ag/AgCl low-profile reference electrode and electrochemical-grade graphite carbon rod of 6.4 mm diameter were acquired from Pine Research. The surface area of contact on the graphite rod with the reaction solution for electrochemical oxidation was approximately 8.7 cm² for all runs.

For each experimental trial, 16.5 ml of total reaction solution consisting of Tetrabutylammonium Nickel (II) Maleonitriledithiolate, lithium nitrate, 1-pentene, and acetonitrile was prepared to appropriate target concentration. The acetonitrile was saturated with lithium nitrate and mixed well for 24 hours with a stirrer. The mixed solution was filtered through 0.22 μm syringe filters to be then used as the solvent to dissolve Tetrabutylammonium Nickel (II) Maleonitriledithiolate and 1-pentene. Before applying the oxidative current in each trial, 500 μl of the solution was mixed with 100 μl of 99.99% deuterated acetonitrile (purchased from Fisher Scientific) and stored in a 5 mm diameter NMR tube for analysis later. An additional 20 μl of the solution was mixed with an appropriate amount of pure acetonitrile to a constant target diluted concentration of 5 μM [Ni(mnt)₂]ⁿ for UV-Vis analysis purpose (i.e., to avoid oversaturation of spectra measurement). For example, 20 μl of 5 mM [Ni(mnt)₂]ⁿ was be diluted with 20 ml of pure acetonitrile in a scintillation vial and then transferred to a UV-Vis quartz-glass cuvette for spectral measurement.

Cleaned reactor parts were assembled prior to each trial with a new Nafion-117 membrane installed. 10 ml of 1 M sulfuric acid was used in the counter-side of the cell while the remaining reaction solution was carefully added into the working-side of the reactor. The reactor was then tightly sealed with a chemical-resistant Viton gasket and a plastic lid. Biologic Potentiostat VSP-300 was set up to connect with a 3-electrode configuration to carefully control the current delivered to the cell. Experiments were conducted in ambient temperature and pressure of approximately 25°C and 1 atm, respectively, while small stir bars rotating at 500 rpm constantly mixed both chambers during reaction. A constant current density was applied for predetermined t_rxn in each trial depending on the amount of initial [Ni(II)(mnt)₂]²⁻ present in the reaction solution. After oxidation, another 500 μl and 20 μl, were extracted and prepared for NMR and UV-Vis spectroscopy, respectively in the same manner. For H-NMR analysis, the tetrabutylammonium ion acted as the reference signal for quantification purposes. The Cary 60 UV-Vis and the Bruker AVANCE NEO 500 MHz NMR were used to characterize the liquid sample. For the current density experiments depicted in Figure 6, constant currents of 1, 2, 3, 5 and 7 mA were applied to identical electrode setups. Higher current was unable to be applied due to potential overloads caused by the relatively high resistance of the organic electrolyte used in the electrochemical cell.

In the time-dynamic study, 4 combinations of initial concentration conditions indicated that equilibration of the complexation reaction can be reached in approximately 72 hours after electrochemical oxidation step. Figure 7 displays the complex concentration profile for the experiment with initial concentrations of 4 mM [Ni(mnt)₂]ⁿ and 8 mM 1-pentene. The other 3 initial concentration conditions are shown in Section S4 in supporting information: (1) 2 mM [Ni(mnt)₂]ⁿ and 2 mM 1-pentene, (2) 4 mM [Ni(mnt)₂]ⁿ and 2 mM 1-pentene, and (3) 4 mM [Ni(mnt)₂]ⁿ and 12 mM 1-pentene. A concentrated stock solution containing the oxidized [Ni(II)(mnt)₂]²⁻ species without olefin was prepared beforehand and was mixed with appropriate amount of 1-pentene, LiNO₃ saturated acetonitrile, and deuterated acetonitrile in the NMR tube. This preparation method allowed better tracking of elapsed time of complexation reaction. 16ml of 20 mM [Ni(IV)(mnt)₂]⁰ starting stock acetonitrile solution, saturated with LiNO₃ supporting electrolyte and absent of olefin, was prepared by applying 0.57 mA/cm² oxidative current density for approximately 3.5 hours. The sample solution also consisted of 10% by volume deuterated acetonitrile when undergoing H-NMR spectroscopy measurement with the Bruker AVANCE NEO 500 MHz NMR.